A Proposed Model for Customer Churn Prediction and Factor Identification Behind Customer Churn in Telecom Industry

Abstract

AbstractIn the telecommunication area, an immense bulk of data is being created consistently because of a huge customer base. Chiefs and business analyst stressed that accomplishing new customer is prohibitive instead of holding the present ones. Business specialists and CRM analyzers need to know the explanations behind customer’s attrition. This paper profound a model for predicting customer churn that utilizes numerous machine learning classification algorithms, as well as the factors and reasons behind customer churn in the telecommunication sector, and the proposed customer churn prediction model initially classifies customer’s data, utilizing classification algorithms, in which the gradient boost algorithm carried out well with 96% accurately grouped cases. In addition, the proposed customer churn prediction model distinguished customer churn factors that are basic in deciding the underlying drivers of customer churn in which the attribute selected classifier has been utilized from Weka tool, and by realizing the critical customer churn factors from customers, CRM can increase their profitability and prescribe pertinent advancements to the likely churn customers. The proposed customer churn prediction model is assessed utilizing measurement matrices, for example, accuracy, precision, recall, F-measure, and receiving operating characteristics (ROCs) area. The outcomes uncover that the proposed customer churn prediction model delivered better churn classification utilizing the gradient boost algorithm and factors behind customer churn, utilizing attribute selected classifier algorithm.KeywordsChurn predictionTelecomCRMBig dataMachine learningData mining techniquesFactor identification